Cooperative Estimation to Reconstruct the Time-varying Flow Field Using Multiple AUVs

We investigate the problem of recovering the flow field of time-varying parameters from the motion data of sensors of an autonomous underwater vehicle (AUV) in this paper. We develop a parametric flow model based on the concept of incompressibility to provide physical theoretical support. Then, considering the influence of the unknown flow field on the AUVs trajectory, we define: 1) the deviation between the actual relative position and the predicted relative position between each vehicle and its neighbors are regarded as the relative motion integral error, which can be obtained by local measurement; 2) The deviation between the actual position and the predicted position of each vehicle is regarded as the absolute motion integral error, which can be obtained by the global positioning system (GPS). We establish the relationship between the motion integration error and the time-varying flow field, and formulate it as a set of nonlinear equations, a row-wise iterative algorithm is proposed to estimate the time-varying flow field. Finally, the effectiveness of the algorithm is verified by simulation.